CN109063737A - Image processing method, device, storage medium and mobile terminal - Google Patents
Image processing method, device, storage medium and mobile terminal Download PDFInfo
- Publication number
- CN109063737A CN109063737A CN201810714996.0A CN201810714996A CN109063737A CN 109063737 A CN109063737 A CN 109063737A CN 201810714996 A CN201810714996 A CN 201810714996A CN 109063737 A CN109063737 A CN 109063737A
- Authority
- CN
- China
- Prior art keywords
- image
- intersection
- cluster
- similarity
- preset threshold
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
- G06F18/232—Non-hierarchical techniques
- G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
Abstract
This application involves a kind of image processing method, device, storage medium and mobile terminals.The processing method includes: to extract the characteristics of image of image to be processed；The similarity between each described image feature is obtained, the image to be processed that the similarity of described image feature is greater than similarity threshold is polymerized to one kind, obtains the cluster intersection of the image to be processed；When the clustering information of the first cluster intersection meets default adjusting condition, the similarity threshold is adjusted to carry out multiple clustering processing to the image in the first cluster intersection；Image intersection after obtaining multiple clustering processing.Above-mentioned image processing method makes the image classification clear layer in image intersection by repeatedly cluster, and compound with regular structure facilitates user to select image, improves user experience.
Description
Technical field
This application involves computer application fields, more particularly to a kind of image processing method, device, storage medium and shifting
Dynamic terminal.
Background technique
Currently, more and more users are clapped using intelligent electronic device with the rapid development of intelligent electronic device
According to.Intelligent electronic device can to shooting obtain image classify, it is more and more however as the image of acquisition, to image into
The classification of row classification and the quantity of each classification also increase, and cause user to take time and effort when selecting image, reduce user's body
Degree of testing.
Summary of the invention
The embodiment of the present application provides a kind of image processing method, device, storage medium and mobile terminal, can be convenient user
Image is selected, user experience is improved.
A kind of image processing method, comprising:
Extract the characteristics of image of image to be processed；
The similarity between each described image feature is obtained, the similarity is greater than to the image to be processed of similarity threshold
It is polymerized to one kind, obtains the first cluster intersection of the image to be processed；
When the clustering information of the first cluster intersection meets default adjusting condition, the similarity threshold is adjusted with right
Image in the first cluster intersection carries out multiple clustering processing；
Image intersection after obtaining multiple clustering processing.
A kind of processing unit of image, described device include:
Extraction module, for extracting the characteristics of image of image to be processed；
Cluster module, for obtaining the similarity between each described image feature, and by the similarity of described image feature
Image to be processed greater than similarity threshold is polymerized to one kind, obtains the first cluster intersection of the image to be processed；
Adjustment module, when the clustering information for clustering intersection when described first meets default adjusting condition, described in adjusting
Similarity threshold is to carry out multiple clustering processing to the image in the first cluster intersection；
Module is obtained, for obtaining the image intersection after multiple clustering processing.
A kind of storage medium is stored thereon with computer program, realizes this when the computer program is executed by processor
The step of applying for the image processing method in each embodiment.
A kind of mobile terminal, including memory and processor store computer program, the calculating in the memory
When machine program is executed by the processor, so that the step of processor executes the image processing method.
Image processing method, device, storage medium and mobile terminal in the embodiment of the present application, by extracting image to be processed
Characteristics of image；The similarity between each described image feature is obtained, the similarity of described image feature is greater than similarity threshold
The image to be processed of value is polymerized to one kind, obtains the cluster intersection of the image to be processed；When the cluster of the first cluster intersection
When information meets default adjusting condition, it is multiple to carry out to the image in the first cluster intersection to adjust the similarity threshold
Clustering processing；Image intersection after obtaining multiple clustering processing, to make the image classification clear layer in image intersection, structure
It is regular, facilitate user to select image, improves user experience.
Detailed description of the invention
In order to illustrate the technical solutions in the embodiments of the present application or in the prior art more clearly, to embodiment or will show below
There is attached drawing needed in technical description to be briefly described, it should be apparent that, the accompanying drawings in the following description is only this
Some embodiments of application for those of ordinary skill in the art without creative efforts, can be with
It obtains other drawings based on these drawings.
Fig. 1 is the application scenario diagram of image processing method in one embodiment；
Fig. 2 is the flow chart of image processing method in one embodiment；
Fig. 3 is the structural schematic diagram of convolutional neural networks in an embodiment；
Fig. 4 is that similarity threshold is adjusted in an embodiment to carry out multiple clustering processing to the image in the first cluster intersection
Method flow diagram；
Fig. 5 is that similarity threshold is adjusted in another embodiment to carry out at multiple cluster to the image in the first cluster intersection
The method flow diagram of reason；
Fig. 6 is the method flow diagram that the image intersection after multiple clustering processing is obtained in an embodiment；
Fig. 7 is the method flow diagram that the image intersection after multiple clustering processing is obtained in another embodiment；
Fig. 8 is the method flow diagram of image processing method in another embodiment；
Fig. 9 is the structural block diagram of the processing unit of image in one embodiment；
Figure 10 A is the schematic diagram of internal structure of mobile terminal in one embodiment；
Figure 10 B is the schematic diagram of internal structure of server in one embodiment；
Figure 11 is the schematic diagram of image processing circuit in one embodiment.
Specific embodiment
It is with reference to the accompanying drawings and embodiments, right in order to which the objects, technical solutions and advantages of the application are more clearly understood
The application is further elaborated.It should be appreciated that specific embodiment described herein is only used to explain the application, and
It is not used in restriction the application.
Fig. 1 is the application scenario diagram of image processing method in one embodiment.As shown in Figure 1, computer equipment 110 and clothes
Being engaged in device 120 can be by network connection communication.Computer equipment 110 can carry out multiclass classification, above-mentioned multistage to stored image
Classification may include the multiclass classification of same dimension or the multiclass classification of different dimensions.For example, computer equipment can first press image
Classify according to time dimension, then classifies to sorted image collection according to place dimension.Computer equipment 110 can will store figure
As being synchronized to server 120, server 120 can carry out multiclass classification to the image received, and the result of multiclass classification is sent out
Give computer equipment 110.Server 120 also can receive the classification results that computer equipment 110 uploads, according to computer equipment
110 classification results uploaded update the classification to image has been stored.That is computer equipment 110 and server 120 can be to figure
As carrying out multiclass classification, wherein when classification results are synchronous between progress computer equipment 110 and server 120, if to image
Classification be related to the manual operation of user, then user of being subject to be manually operated after classification results.If not to the classification of image
It is related to the manual operation of user, then is subject to the classification results of server 120.
Fig. 2 is the flow chart of image processing method in one embodiment.As shown in Fig. 2, a kind of image processing method, including
Step 202 is to step 208.
Step 202: extracting the characteristics of image of image to be processed.
Computer equipment is input to convolutional neural networks after getting image to be processed, by image to be processed
(Convolutional Neural Network, CNN) model carries out feature extraction.CNN model refers in traditional multilayer mind
A kind of a kind of neural network model for image classification and identification through growing up on the basis of network, it is opposite with it is traditional
Multilayer neural network, CNN model introduce convolution algorithm and pond algorithm.Wherein, refer to will be in regional area for convolution algorithm
Data are weighted a kind of mathematical algorithm of superposition, and pond algorithm, which refers to, carries out the one of sampling processing for the data in regional area
Kind mathematical algorithm.
It is formed specifically, CNN model is replaced by convolutional layer with pond layer, as shown in figure 3,301 input picture of input layer,
Convolutional layer 302 carries out image characteristics extraction, the image of 303 pairs of convolutional layer of pond layer to each regional area of the image of input layer
Feature is sampled to reduce dimension, is then attached again with the full articulamentum 304 of several layers to characteristics of image, with the last layer
The output valve of hidden layer 305 is the feature finally extracted.
Image to be processed can be arbitrary image.The figure of the image or mobile terminal local photograph album such as downloaded from network
Picture or the image of shooting.
Step 204: obtaining the similarity between each characteristics of image, and the similarity of characteristics of image is greater than similarity threshold
Image to be processed be polymerized to one kind, obtain image to be processed first cluster intersection.
Specifically, the characteristics of image that the last layer hidden layer 305 exports is input to classifier, classifier calculated is any
The image that similarity is greater than similarity threshold is polymerized to one kind by clustering algorithm by the similarity of the characteristics of image of two images.
It should be understood that ground, is carried out in similarity calculation, the present embodiment extracts the SIFT of image first based on characteristics of image
(Scale-Invariant Feature Transform, scale invariant feature conversion) feature vector, then calculate any two width figure
The similarity of the SIFT feature vector of picture.For example, being made after the SIFT feature vector of two images generates using Euclidean distance
It is measured for the similarity determination of the SIFT feature vector in two images.If the SIFT feature vector of two images it is European away from
From some threshold value is less than, then illustrate that the similarity of two width figures is higher namely the similarity of characteristics of image is greater than some similarity threshold
Value.At this point, this two width figure can be classified as one kind.
It should be noted that SIFT is a kind of algorithm for detecting local feature, the characteristic point which passes through extraction image
The characteristics of image of image is described, for different image types, different characteristic points can also be extracted, such as include portrait
Image type can extract face eigen vector as characteristics of image.In other embodiments, characteristics of image further include cuisines,
The features such as landscape or portrait, are not specifically limited herein.
In the present embodiment, clustering algorithm includes C means clustering algorithm, kmediods clustering algorithm, Self-organizing Maps SOM
Clustering algorithm or fuzzy C-mean algorithm FCM clustering algorithm.
Step 206: when the clustering information of the first cluster intersection meets default adjusting condition, adjusting similarity threshold with right
Image in first cluster intersection carries out multiple clustering processing.
It should be understood that ground, similarity is the clustering measure of image.For example, when the similarity of two images is greater than a certain similar
When spending threshold value, two images are classified as one kind, if increasing similarity threshold, two images will not belong to same class.Therefore, similar
Degree threshold value can measure the quantity of cluster image.
In the present embodiment, the clustering information of the first cluster intersection includes categorical measure and amount of images, which is
Amount of images in of all categories, and similarity threshold and categorical measure proportional are in interior amount of images of all categories
Inversely prroportional relationship.That is, increasing similarity threshold when one timing of quantity of image to be processed, clustering the categorical measure in intersection
It will increase, and interior amount of images of all categories can reduce.Therefore, by adjusting similarity threshold, the first cluster intersection can be made
In categorical measure and interior amount of images control of all categories needs range.
It should be understood that ground, the categorical measure for clustering intersection when first and interior amount of images of all categories be not in the range of needs
When interior, the clustering information of the as first cluster intersection meets default adjusting condition, at this time can be by adjusting similarity threshold with right
Image in the first cluster intersection carries out multiple clustering processing, so that categorical measure in the first cluster intersection and of all categories
Range of the interior amount of images control in needs.
In one embodiment, default adjusting condition includes any in the following conditions:
Categorical measure is less than the second preset threshold and the other amount of images of any sort is greater than the first preset threshold.This condition is
Within the limits prescribed, and there are the other amount of images of any sort not in defined range for categorical measure in first cluster intersection
It is interior.
Categorical measure is greater than the second preset threshold.This condition is that the categorical measure in the first cluster intersection exceeds defined model
It encloses.
Step 208: the image intersection after obtaining multiple clustering processing.
In one embodiment, according to user demand, similarity threshold can be preset, similarity threshold is preset by this and is obtained
Similarity threshold is adjusted when the clustering information of the first cluster intersection meets default adjusting condition to the first cluster intersection.It is opening
Begin before adjusting similarity threshold, presets the pre-set categories quantity and interior pre-set image quantity of all categories in clustering information.
During adjusting similarity threshold, the categorical measure and amount of images of all categories in the first cluster intersection are with similarity threshold
It is worth dynamic change, when categorical measure reaches pre-set categories quantity, stops adjusting similarity threshold, obtain the figure after adjusting at this time
As intersection, namely the repeatedly image intersection after clustering processing.The image intersection includes the first image intersection and the second image intersection,
First image intersection is the classification intersection that each amount of images is less than pre-set image quantity；Second image intersection is greater than for amount of images
Any classification of pre-set image quantity.
It should be noted that the first image intersection is to meet the image intersection namely photograph album of user demand；Second image closes
Collection can be used as individual photograph album also to need the image intersection clustered again.
In above-mentioned image processing method, by the characteristics of image for extracting image to be processed；Obtain each described image feature it
Between similarity, the image to be processed that the similarity of described image feature is greater than similarity threshold is polymerized to one kind, is obtained described
The cluster intersection of image to be processed；When the clustering information of the first cluster intersection meets default adjusting condition, described in adjusting
Similarity threshold is to carry out multiple clustering processing to the image in the first cluster intersection；Figure after obtaining multiple clustering processing
As intersection, to make the image classification clear layer in image intersection, compound with regular structure facilitates user to select image, improves use
Family Experience Degree.
Fig. 4 is that similarity threshold is adjusted in an embodiment to carry out multiple clustering processing to the image in the first cluster intersection
Method flow diagram, as shown in figure 4, the method comprising the steps of 402 to step 404.
Step 402: obtaining the categorical measure in adjustment process and amount of images of all categories.
Step 404: when categorical measure is less than the second preset threshold and the other amount of images of any sort is greater than the first default threshold
When value, increase similarity threshold until categorical measure reaches the second preset threshold.
Specifically, due to the categorical measure proportional in similarity threshold and cluster intersection, and it is of all categories interior
Amount of images be in inversely prroportional relationship.That is, increasing similarity threshold when one timing of quantity of image to be processed, clustering intersection
In categorical measure will increase, and it is of all categories in amount of images can reduce.Therefore cluster can be preset according to user demand
The categorical measure upper limit in intersection (photograph album) is the second preset threshold and the interior amount of images upper limit of all categories is first default
Threshold value.Image to be processed is being carried out in cluster process using clustering algorithm (similarity to be greater than similarity threshold
Image to be processed is polymerized to one kind), real-time detection interior amount of images of all categories.When detecting that the amount of images in a certain classification is super
When crossing amount of images the first preset threshold of the upper limit, on the one hand increasing similarity threshold reduces the amount of images in the category；Separately
On the one hand, the categorical measure that increasing similarity threshold can make to cluster in intersection increases.Therefore it according to user demand, presets poly-
The categorical measure upper limit in class intersection (photograph album) is the second preset threshold, when category quantity reaches the second preset threshold of the upper limit
When, stop the increase to similarity threshold.
It should be noted that needing first to detect the categorical measure and image of all categories in the first cluster intersection before adjusting
Quantity.First cluster intersection is the image intersection arrived clustered according to initial similarity threshold to image to be processed.When
When the clustering information (categorical measure and amount of images of all categories) of one cluster intersection meets default adjusting condition, similarity is adjusted
Threshold value is to carry out multiple clustering processing to the image in the first cluster intersection.
In the present embodiment, the second preset threshold of the categorical measure upper limit and the interior figure of all categories in intersection are clustered by setting
As the first preset threshold of the upper limit of the number, so that categorical measure and amount of images control within the scope of preset quantity, to reduce
User selects the time of image, improves user experience.
Fig. 5 is that similarity threshold is adjusted in another embodiment to carry out at multiple cluster to the image in the first cluster intersection
The method flow diagram of reason, as shown in figure 5, this method further includes step 406.
Step 406: when categorical measure is greater than the second preset threshold, reducing similarity threshold until categorical measure is less than the
Two preset thresholds.
Wherein, the second preset threshold is to cluster the upper limit value of the categorical measure in intersection.Specifically, due to similarity threshold
Value and the categorical measure proportional in cluster intersection, are in inversely prroportional relationship with interior amount of images of all categories.That is, when to
The timing of quantity one for handling image, increases similarity threshold, and the categorical measure clustered in intersection will increase, and it is of all categories in
Amount of images can reduce.Therefore can be according to user demand, presetting the categorical measure upper limit in cluster intersection (photograph album) is the
Two preset thresholds and the interior amount of images upper limit of all categories are the first preset threshold.In use clustering algorithm to figure to be processed
As carrying out in cluster process and (image to be processed that the similarity is greater than similarity threshold being polymerized to one kind), real-time detection is each
Amount of images in classification.When detect cluster intersection in categorical measure be more than the second preset threshold of the categorical measure upper limit
When, reduce similarity threshold so that the categorical measure in cluster intersection declines；When categorical measure drops to the second preset threshold,
Stop reducing similarity threshold.
In the present embodiment, categorical measure control is made by the second preset threshold of the categorical measure upper limit in setting cluster intersection
System is within the scope of preset quantity, and the categorical measure avoided in cluster intersection is various, and the time that user selects image longer asks
Topic.
Fig. 6 is the method flow diagram that the image intersection after multiple clustering processing is obtained in an embodiment, as shown in fig. 6, should
Method includes step 602 to step 608.
Step 602: obtaining the second cluster intersection generated after multiple clustering processing.
Step 604: the amount of images of all categories in detection the second cluster intersection.
Step 606: in the second cluster intersection, all categories that amount of images is less than or equal to the first preset threshold being gathered
It is combined into the first image intersection.
Step 608: in the second cluster intersection, amount of images being greater than any classification of the first preset threshold as second
Image intersection.
In the present embodiment, when due to adjusting similarity threshold, the categorical measure and interior image of all categories in intersection are clustered
Quantity changes simultaneously, therefore, when amount of images be greater than the first preset threshold when, can categorical measure allow quantitative range (i.e.
Second preset threshold) in similarity threshold is adjusted so that the amount of images in classification reduces.
It should be understood that ground, is adjusted similarity threshold in the quantitative range (i.e. the second preset threshold) that categorical measure allows
Value, the interior amount of images of all categories after adjusting includes two kinds of situations: one is be greater than the first preset threshold；Another kind is less than
Or it is equal to the first preset threshold.Therefore, the amount of images in of all categories is detected after adjusting, is obtained amount of images according to demand and is less than
Or all categories equal to the first preset threshold, and it is polymerized as the first image intersection, and to be greater than first default for amount of images
The classification of threshold value is separately as the second image intersection.
Specifically, obtaining the second cluster intersection after to the multiple clustering processing of image in the first cluster intersection.?
In two cluster intersections, categorical measure meets preset condition, i.e., categorical measure is less than the second preset threshold.At this point, detection second
The amount of images of all categories in intersection is clustered, all categories that amount of images is less than or equal to the first preset threshold are polymerized to
First image intersection；Amount of images is greater than any classification of the first preset threshold as the second image intersection.
It should be noted that the first image intersection includes the class that all amount of images are less than or equal to the first preset threshold
Not, these classifications are condensed together as a photograph album, and the second image intersection is that amount of images is greater than the first preset threshold
Any classification, be a classification, using the second image intersection as individual photograph album.It according to demand can be by the second image intersection
It clusters, i.e., by increase similarity threshold, the second image intersection is clustered again again, the classification of the second image intersection is thin
Point.For example, the second image intersection that classification is cuisines is clustered again so that the classification of the second image intersection is cuisines as an example
The image intersection of the classifications such as hamburger, noodles, fruit can be formed after (increasing similarity threshold).In the present embodiment, by second
The cluster again of image intersection, the amount of images avoided in same category is more, and user browses inconvenient problem.To image
Repeatedly cluster, so that the mode classification of image more refines, is conducive to the image that user disaggregatedly checks corresponding types.
Fig. 7 is the method flow diagram that the image intersection after multiple clustering processing is obtained in another embodiment, as shown in fig. 7,
The method comprising the steps of 602 to step 608, further includes step 605.
Step 605: in the second cluster intersection, all categories that amount of images is less than third predetermined threshold value being merged into one
Class.Wherein, third predetermined threshold value is less than the first preset threshold.
Specifically, third predetermined threshold value can be used as the lower limit value of the amount of images in of all categories, when detecting multiple classes
When amount of images in not is less than third predetermined threshold value, illustrate the image quantity of these classifications, it can be by the figure of these classifications
As merging into one kind, the classification information after merging can be labeled as other classifications.It should be noted that third predetermined threshold value is less than the
One preset threshold, i.e. third predetermined threshold value can be used as the lower limit value of the amount of images in classification, and the first preset threshold can be used as class
The upper limit value of amount of images in not.
It should be noted that device merges after the categories combination that multiple images quantity is less than third predetermined threshold value is a kind of
The amount of images in classification (other classifications) afterwards can be greater than the first preset threshold, might be less that or be equal to the first default threshold
Value.If being less than or equal to the first preset threshold, as one kind in the first image intersection；If more than the first default threshold
Value, then by it separately as the second image intersection.
In the present embodiment, is merged, avoided first by the classification for being less than third predetermined threshold value to amount of images
In image intersection because of the image quantity in a certain classification, cause categorical measure various, user browses inconvenient problem.It is right
The less classification of image merges, so that the mode classification of image is more regular, is conducive to improve user experience.
Fig. 8 is the method flow diagram of image processing method in another embodiment, as shown in figure 8, in the second cluster intersection,
After amount of images is greater than any classification of the first preset threshold as the second image intersection, including step 802 is to step
804。
Step 802: adjusting similarity threshold, multiple clustering processing is carried out to the image in the second image intersection.
Step 804: obtaining the third cluster intersection for meeting default cluster condition after multiple clustering processing.Wherein, described pre-
If cluster condition include: third cluster intersection in categorical measure less than the second preset threshold and the other amount of images of any sort it is small
In the first preset threshold.
It should be noted that after obtaining the second image intersection, it is possible to increase similarity threshold is in the second image intersection
Image is repeatedly clustered, and the classification of the second image intersection is segmented, and the amount of images avoided in same category is more.?
The third that can obtain meeting cluster condition after repeatedly clustering cluster intersection is carried out, i.e. categorical measure in third cluster intersection is less than
The second preset threshold and other amount of images of any sort is less than the first preset threshold.It should be appreciated that clustering intersection in third
In be likely present amount of images be greater than the first preset threshold classification, the category can as individual image intersection, and
Cluster subdivision is carried out to it, specific method is identical as the clustering method to the second image intersection, and details are not described herein.
In above-mentioned image processing method, computer equipment can carry out image to be processed according to similarity threshold size multistage
Cluster.For example, to image clustering to be processed can be regarded as level-one cluster (i.e. by the similarity be greater than similarity threshold wait locate
Reason image is polymerized to one kind)；Second level cluster can be regarded as to the cluster that the second image intersection carries out.Except level-one cluster and second level cluster
Outside, computer equipment can also carry out three-level cluster to image to be processed, level Four clusters N grades of clusters etc..Wherein, level-one is poly-
The range of class is greater than second level cluster, the range of second level cluster is clustered greater than three-level, and so on, i.e. previous stage cluster is rear stage
The father of cluster clusters, and rear stage cluster is the son cluster of previous stage cluster.It is corresponding that clusters at different levels are prestored in computer equipment
Clustering rule can obtain the level-one clustering rule prestored when computer equipment carries out level-one cluster to image to be processed.Its
In, clustering rules at different levels are corresponding with similarity threshold.For example, clustering rule can be to be polymerized to image to be processed by similarity
The image that different poly groups, i.e. similarity are greater than similarity threshold, which is drawn, gathers same group.
Computer equipment can adjust the similarity threshold poly group (i.e. second image intersection) more to above-mentioned amount of images into
Row second level cluster.After carrying out level-one cluster to image to be processed, level-one can be obtained and cluster corresponding first image intersection and the
Two image intersections, wherein the second image intersection is the more poly group of amount of images；First image intersection is image quantity
Poly group set.Computer equipment is when carrying out secondary cluster to the second image intersection, it is possible to increase similarity threshold is (relative to level-one
For cluster), and according to default clustering rule (draw the image that similarity is greater than similarity threshold and gather same group) to the
Two image intersections carry out second level cluster.
In one embodiment, default clustering rule can the classification information of the according to second image intersection the second image is closed
Collection is divided into different dimensions.Specifically, computer equipment can obtain the second figure when clustering to the second image intersection second level
As the classification information of intersection, and the dimension of second level cluster is carried out to the second image collection, according to the dimension to the second image set
It closes and carries out second level cluster.The classification information of above-mentioned second image intersection may include cuisines, landscape, portrait, animal and other classifications.
Wherein, the dimension of second level cluster is related to the classification information of the second image intersection, i.e., the dimension of second level cluster is the conjunction of the second image
Thin poly- direction on the basis of the classification information of collection.For example, using classification information as animal category, it is poly- to the second level of the second image intersection
The dimension of class is the corresponding species of animal, such as cat, dog, fish.Dimension to the second level cluster of landscape classification can be blue sky, grass
Ground, sandy beach etc.；Dimension to the second level cluster of portrait classification can be the corresponding gender of portrait, age etc.；To the two of cuisines classification
The dimension of grade cluster can be the corresponding food species of cuisines.
Fig. 9 is the structural block diagram of the processing unit of image in one embodiment.As shown in figure 9, a kind of processing of image fills
It sets, including extraction module 910, cluster module 920, adjustment module 930 and acquisition module 940.Wherein:
Extraction module 910, for extracting the characteristics of image of image to be processed.
Cluster module 920, for obtaining the similarity between each described image feature, and by the similar of described image feature
The image to be processed that degree is greater than similarity threshold is polymerized to one kind, obtains the first cluster intersection of the image to be processed.
Adjustment module 930, for adjusting institute when the clustering information of the first cluster intersection meets default adjusting condition
Similarity threshold is stated to carry out multiple clustering processing to the image in the first cluster intersection.
Module 940 is obtained, for obtaining the image intersection after multiple clustering processing.
In the present embodiment, extraction module 910 extracts the figure of image to be processed based on the convolutional neural networks model trained
As feature；Cluster module 920 obtains the similarity between each described image feature, and uses clustering algorithm by described image feature
Similarity be greater than similarity threshold image to be processed be polymerized to one kind, obtain the image to be processed first cluster intersection；
Adjustment module 930 is used to adjust the similarity when the clustering information of the first cluster intersection meets default adjusting condition
Threshold value is to carry out multiple clustering processing to the image in the first cluster intersection；It obtains module 940 and obtains multiple clustering processing
Image intersection afterwards.Above-mentioned image processing apparatus makes the image classification clear layer in image intersection, structure by repeatedly cluster
It is regular, facilitate user to select image, improves user experience.
In one embodiment, adjustment module 930, comprising:
First acquisition unit, for obtaining categorical measure and amount of images of all categories in adjustment process.
First adjusts unit, for the categorical measure less than the second preset threshold and the other amount of images of any sort is greater than
When the first preset threshold, increase the similarity threshold until the categorical measure reaches the second preset threshold.
In one embodiment, adjustment module 930, further includes:
Second adjusts unit, for reducing the similarity threshold when the categorical measure is greater than the second preset threshold
Until the categorical measure is less than the second preset threshold.
In one embodiment, module 940 is obtained, comprising:
Second acquisition unit, for obtaining the second cluster intersection generated after the multiple clustering processing.
Detection unit, for detecting the amount of images of all categories in the second cluster intersection.
Described image quantity is less than or equal to the first default threshold in the second cluster intersection by the first polymerized unit
The all categories of value are polymerized to the first image intersection
Described image quantity is greater than appointing for the first preset threshold in the second cluster intersection by the second polymerized unit
One classification is as the second image intersection.
In one embodiment, module 940 is obtained, further includes:
Described image quantity is less than the institute of third predetermined threshold value in the second cluster intersection by third polymerization unit
There is categories combination for one kind；The third predetermined threshold value is less than first preset threshold.
In one embodiment, a kind of processing unit of image, further includes:
Second level cluster module carries out the image in the second image intersection more for adjusting the similarity threshold
Secondary clustering processing；Obtain the third cluster intersection for meeting default cluster condition after multiple clustering processing；The default cluster condition
Include: in third cluster intersection categorical measure is less than the second preset threshold and the other amount of images of any sort is less than first
Preset threshold.
Although should be understood that each step in the flow chart of Fig. 2, Fig. 4, Fig. 5, Fig. 6, Fig. 7, Fig. 8 according to arrow
Instruction successively show that but these steps are not that the inevitable sequence according to arrow instruction successively executes.Unless having herein
Explicitly stated, there is no stringent sequences to limit for the execution of these steps, these steps can execute in other order.And
And at least poly- step in Fig. 2, Fig. 4, Fig. 5, Fig. 6, Fig. 7, Fig. 8 may include multiple sub-steps or multiple stages, this
A little step or stage are not necessarily to execute completion in synchronization, but can execute at different times, this is a little
Step perhaps the stage execution sequence be also not necessarily successively carry out but can be with the sub-step of other steps or other steps
Rapid or the stage at least one poly- to execute in turn or alternately.
Modules draws poly- be only used for for example, in other embodiments, can incite somebody to action in the processing unit of above-mentioned image
The processing unit of image is drawn as required gathers for different modules, to complete the poly- function of whole or portion of the processing unit of above-mentioned image
Energy.
The embodiment of the present application also provides a kind of mobile terminal.The mobile terminal includes memory and processor, the memory
In store computer program, the computer program by the processor execute when so that the processor executes the image processing method
The step of method.
The mobile terminal includes memory and processor, stores computer program in the memory, the computer program
When being executed by the processor, so that the step of processor executes the image processing method.
The embodiment of the present application also provides a kind of storage medium.A kind of storage medium, is stored thereon with computer program, the meter
The step of calculation machine program realizes the image processing method when being executed by processor.
Figure 10 A is the schematic diagram of internal structure of mobile terminal in one embodiment.As shown in Figure 10 A, the mobile terminal packet
Include processor, memory and the network interface connected by system bus.Wherein, which calculates and controls energy for providing
Power supports the operation of entire mobile terminal.Memory for storing data, program etc., at least one calculating is stored on memory
Machine program, the computer program can be executed by processor, to realize that is provided in the embodiment of the present application is suitable for mobile terminal
Wireless network communication method.Memory may include non-volatile memory medium and built-in storage.Non-volatile memory medium storage
There are operating system and computer program.The computer program can be performed by processor, for realizing above each embodiment
A kind of provided image processing method.Built-in storage provides for the operating system computer program in non-volatile memory medium
The running environment of cache.Network interface can be Ethernet card or wireless network card etc., for external mobile terminal into
Row communication.The mobile terminal can be mobile phone, tablet computer or personal digital assistant or wearable device etc..
Figure 10 B is the schematic diagram of internal structure of server (or cloud etc.) in one embodiment.As shown in Figure 10 B, the clothes
Business device includes processor, non-volatile memory medium, built-in storage and the network interface connected by system bus.Wherein, should
Processor supports the operation of entire mobile terminal for providing calculating and control ability.Memory for storing data, program
Deng, at least one computer program is stored on memory, which can be executed by processor, with realize the application implement
The wireless network communication method suitable for mobile terminal provided in example.Memory may include non-volatile memory medium and memory
Reservoir.Non-volatile memory medium is stored with operating system and computer program.The computer program can performed by processor,
For realizing a kind of image processing method provided by above each embodiment.Built-in storage is in non-volatile memory medium
Operating system computer program provide cache running environment.Network interface can be Ethernet card or wireless network card
Deng for being communicated with external mobile terminal.Server can be formed with the either multiple servers of independent server
Server cluster realize.It will be understood by those skilled in the art that structure shown in Figure 10 B, only with the application side
The block diagram of the relevant portion's poly structure of case, does not constitute the restriction for the server being applied thereon to application scheme, specifically
Server may include perhaps combining certain components or with different components than more or fewer components as shown in the figure
Arrangement.
The realization of modules in the recommendation apparatus of the application program provided in the embodiment of the present application can be computer journey
The form of sequence.The computer program can be run on mobile terminal or server.The program module that the computer program is constituted can
It is stored on the memory of mobile terminal or server.When the computer program is executed by processor, the embodiment of the present application is realized
Described in method the step of.
A kind of computer program product comprising instruction, when run on a computer, so that computer executes image
Processing method.
The embodiment of the present application also provides a kind of mobile terminal.It include image processing circuit in above-mentioned mobile terminal, at image
Reason circuit can use hardware and or software component realization, it may include define ISP (Image Signal Processing, figure
As signal processing) the various processing units of pipeline.Figure 11 is the schematic diagram of image processing circuit in one embodiment.Such as Figure 11 institute
Show, for purposes of illustration only, only showing the various aspects of image processing techniques relevant to the embodiment of the present application.
As shown in figure 11, image processing circuit includes ISP processor 1140 and control logic device 1150.Imaging device 1110
The image data of capture is handled by ISP processor 1140 first, and ISP processor 1140 carries out coazevation to image data can to capture
Image statistics for determining and/or imaging device 1110 one or more control parameters.Imaging device 1110 can wrap
Include the camera with one or more lens 1112 and imaging sensor 1114.Imaging sensor 1114 may include colour filter
Array (such as Bayer filter), imaging sensor 1114 can obtain the light captured with each imaging pixel of imaging sensor 1114
Intensity and wavelength information, and the one group of raw image data that can be handled by ISP processor 1140 is provided.1120 (such as top of sensor
Spiral shell instrument) parameter (such as stabilization parameter) of the image procossing of acquisition can be supplied to ISP processing based on 1120 interface type of sensor
Device 1140.1120 interface of sensor can use SMIA, and (Standard Mobile Imaging Architecture, standard are moved
Dynamic Imager Architecture) interface, other serial or parallel camera interfaces or above-mentioned interface combination.
In addition, raw image data can also be sent to sensor 1120 by imaging sensor 1114, sensor 1120 can base
Raw image data is supplied to ISP processor 1140 or sensor 1120 for original graph in 1120 interface type of sensor
As data storage is into video memory 1130.
ISP processor 1140 handles raw image data pixel by pixel in various formats.For example, each image pixel can
Bit depth with 8,10,14 or 14 bits, ISP processor 1140 can carry out raw image data at one or more images
Reason operation, statistical information of the collection about image data.Wherein, image processing operations can be by identical or different bit depth precision
It carries out.
ISP processor 1140 can also receive image data from video memory 1130.For example, 1120 interface of sensor will be former
Beginning image data is sent to video memory 1130, and the raw image data in video memory 1130 is available to ISP processing
Device 1140 is for processing.Video memory 1130 can be only in poly-, the storage equipment or mobile terminal of memory device
Vertical private memory, and may include DMA (Direct Memory Access, direct direct memory access (DMA)) feature.
1114 interface of imaging sensor is come from or from 1120 interface of sensor or from video memory when receiving
When 1130 raw image data, ISP processor 1140 can carry out one or more image processing operations, such as time-domain filtering.Place
Image data after reason can be transmitted to video memory 1130, to carry out other processing before shown.ISP processor
1140 from video memory 1130 receive processing data, and to the processing data carry out original domain in and RGB and YCbCr color
Image real time transfer in space.Treated that image data may be output to display 1170 for ISP processor 1140, for user
It watches and/or is further processed by graphics engine or GPU (Graphics Processing Unit, graphics processor).In addition,
The output of ISP processor 1140 also can be transmitted to video memory 1130, and display 1170 can be read from video memory 1130
Take image data.In one embodiment, video memory 1130 can be configured to realize one or more frame buffers.This
Outside, the output of ISP processor 1140 can be transmitted to encoder/decoder 1160, so as to encoding/decoding image data.Coding
Image data can be saved, and decompress before being shown in 1170 equipment of display.Encoder/decoder 1160 can be by
CPU or GPU or coprocessor are realized.
The statistical data that ISP processor 1140 determines, which can be transmitted, gives control logic device Unit 1150.For example, statistical data can
It is passed including the images such as automatic exposure, automatic white balance, automatic focusing, flicker detection, black level compensation, 1112 shadow correction of lens
1114 statistical information of sensor.Control logic device 1150 may include execute one or more routines (such as firmware) processor and/or
Microcontroller, one or more routines can statistical data based on the received, determine at control parameter and the ISP of imaging device 1110
Manage the control parameter of device 1140.For example, the control parameter of imaging device 1110 may include that 1120 control parameter of sensor (such as increases
Benefit, the accumulation time of spectrum assignment, stabilization parameter etc.), camera flash control parameter, 1112 control parameter of lens it is (such as poly-
Burnt or zoom focal length) or these parameters combination.ISP control parameter may include for automatic white balance and color adjustment (example
Such as, RGB processing during) 1112 shadow correction parameter of gain level and color correction matrix and lens.
The above are realize above-mentioned image processing method with image processing techniques in Figure 11.
Any reference to memory, storage, database or other media used in this application may include non-volatile
And/or volatile memory.Suitable nonvolatile memory may include read-only memory (ROM), programming ROM (PROM),
Electrically programmable ROM (EPROM), electrically erasable ROM (EEPROM) or flash memory.Volatile memory may include arbitrary access
Memory (RAM), it is used as external cache.By way of illustration and not limitation, RAM is available in many forms, such as
It is static RAM (SRAM), dynamic ram (DRAM), synchronous dram (SDRAM), double data rate sdram (DDR SDRAM), enhanced
SDRAM (ESDRAM), synchronization link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM
(RDRAM), direct memory bus dynamic ram (DRDRAM) and memory bus dynamic ram (RDRAM).
The several embodiments of the application above described embodiment only expresses, the description thereof is more specific and detailed, but simultaneously
The limitation to the application the scope of the patents therefore cannot be interpreted as.It should be pointed out that for those of ordinary skill in the art
For, without departing from the concept of this application, various modifications and improvements can be made, these belong to the guarantor of the application
Protect range.Therefore, the scope of protection shall be subject to the appended claims for the application patent.
Claims (10)
1. a kind of image processing method characterized by comprising
Extract the characteristics of image of image to be processed；
The similarity between each described image feature is obtained, and the image to be processed that the similarity is greater than similarity threshold is gathered
At one kind, the first cluster intersection of the image to be processed is obtained；
When the clustering information of the first cluster intersection meets default adjusting condition, the similarity threshold is adjusted to described
Image in first cluster intersection carries out multiple clustering processing；
Image intersection after obtaining multiple clustering processing.
2. the method according to claim 1, wherein the clustering information includes categorical measure and amount of images,
The default adjusting condition includes any in the following conditions:
The categorical measure is less than the second preset threshold and the other amount of images of any sort is greater than the first preset threshold；
The categorical measure is greater than the second preset threshold.
3. according to the method described in claim 2, it is characterized in that, described adjust the similarity threshold with poly- to described first
Image in class intersection carries out multiple clustering processing, comprising:
Obtain the categorical measure in adjustment process and amount of images of all categories；
When the categorical measure is less than the second preset threshold and the other amount of images of any sort is greater than the first preset threshold, increase
The similarity threshold is until the categorical measure reaches the second preset threshold.
4. according to the method described in claim 3, it is characterized in that, described adjust the similarity threshold with poly- to described first
Image in class intersection carries out multiple clustering processing, further includes:
When the categorical measure is greater than the second preset threshold, reduce the similarity threshold until the categorical measure is less than the
Two preset thresholds.
5. the method according to claim 1, wherein the image intersection obtained after multiple clustering processing, packet
It includes:
Obtain the second cluster intersection generated after the multiple clustering processing；
Detect the amount of images of all categories in the second cluster intersection；
In the second cluster intersection, all categories that described image quantity is less than or equal to the first preset threshold are polymerized to
First image intersection；
In the second cluster intersection, described image quantity is greater than any classification of the first preset threshold as the second image
Intersection.
6. according to the method described in claim 5, it is characterized in that, detection described second clusters the image of all categories in intersection
After quantity, further includes:
In the second cluster intersection, all categories that described image quantity is less than third predetermined threshold value are merged into one kind；
The third predetermined threshold value is less than first preset threshold.
7. according to the method described in claim 5, it is characterized in that, described cluster in intersection described second, by described image
Quantity be greater than the first preset threshold any classification as the second image intersection after, comprising:
The similarity threshold is adjusted, multiple clustering processing is carried out to the image in the second image intersection；
Obtain the third cluster intersection for meeting default cluster condition after multiple clustering processing；The default cluster condition includes:
In third cluster intersection categorical measure is less than the second preset threshold and the other amount of images of any sort is less than first
Preset threshold.
8. a kind of processing unit of image, which is characterized in that described device includes:
Extraction module, for extracting the characteristics of image of image to be processed；
Cluster module is greater than for obtaining the similarity between each described image feature, and by the similarity of described image feature
The image to be processed of similarity threshold is polymerized to one kind, obtains the first cluster intersection of the image to be processed；
Adjustment module, for adjusting described similar when the clustering information of the first cluster intersection meets default adjusting condition
Threshold value is spent to carry out multiple clustering processing to the image in the first cluster intersection；
Module is obtained, for obtaining the image intersection after multiple clustering processing.
9. a kind of storage medium, is stored thereon with computer program, which is characterized in that realized such as when the program is executed by processor
Any image processing method in claim 1 to 7.
10. a kind of mobile terminal including memory, processor and stores the calculating that can be run on a memory and on a processor
Machine program, which is characterized in that the processor is realized as described in any in claim 1 to 7 when executing the computer program
Image processing method.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810714996.0A CN109063737A (en) | 2018-07-03 | 2018-07-03 | Image processing method, device, storage medium and mobile terminal |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810714996.0A CN109063737A (en) | 2018-07-03 | 2018-07-03 | Image processing method, device, storage medium and mobile terminal |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109063737A true CN109063737A (en) | 2018-12-21 |
Family
ID=64818923
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810714996.0A Pending CN109063737A (en) | 2018-07-03 | 2018-07-03 | Image processing method, device, storage medium and mobile terminal |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109063737A (en) |
Cited By (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109800744A (en) * | 2019-03-18 | 2019-05-24 | 深圳市商汤科技有限公司 | Image clustering method and device, electronic equipment and storage medium |
| CN110084267A (en) * | 2019-03-12 | 2019-08-02 | 北京旷视科技有限公司 | Portrait clustering method, device, electronic equipment and readable storage medium storing program for executing |
| CN110245679A (en) * | 2019-05-08 | 2019-09-17 | 北京旷视科技有限公司 | Image clustering method, device, electronic equipment and computer readable storage medium |
| CN110738258A (en) * | 2019-10-16 | 2020-01-31 | Oppo广东移动通信有限公司 | Image classification method and device and terminal equipment |
| CN111062407A (en) * | 2019-10-15 | 2020-04-24 | 深圳市商汤科技有限公司 | Image processing method and device, electronic equipment and storage medium |
| CN111091152A (en) * | 2019-12-17 | 2020-05-01 | 上海云从汇临人工智能科技有限公司 | Image clustering method, system, device and machine readable medium |
| CN111310834A (en) * | 2020-02-19 | 2020-06-19 | 深圳市商汤科技有限公司 | Data processing method and device, processor, electronic equipment and storage medium |
| CN111368880A (en) * | 2020-02-17 | 2020-07-03 | 杭州海康威视系统技术有限公司 | Image clustering method and device, electronic equipment and computer readable storage medium |
| CN111382770A (en) * | 2018-12-29 | 2020-07-07 | 百度在线网络技术（北京）有限公司 | Picture clustering method, device, equipment and storage medium |
| CN111401196A (en) * | 2020-03-10 | 2020-07-10 | 珠海全志科技股份有限公司 | Method, computer device and computer readable storage medium for self-adaptive face clustering in limited space |
| CN111460195A (en) * | 2020-03-26 | 2020-07-28 | Oppo广东移动通信有限公司 | Picture processing method and device, storage medium and electronic equipment |
| CN111680753A (en) * | 2020-06-10 | 2020-09-18 | 创新奇智(上海)科技有限公司 | Data labeling method and device, electronic equipment and storage medium |
| CN111753114A (en) * | 2020-06-04 | 2020-10-09 | 北京字节跳动网络技术有限公司 | Image pre-labeling method and device and electronic equipment |
| CN112767410A (en) * | 2020-12-31 | 2021-05-07 | 北京航空航天大学 | MRI brain tissue clustering segmentation method |
| CN113052245A (en) * | 2021-03-30 | 2021-06-29 | 重庆紫光华山智安科技有限公司 | Image clustering method and device, electronic equipment and storage medium |
| CN113378020A (en) * | 2021-06-08 | 2021-09-10 | 深圳Tcl新技术有限公司 | Acquisition method, device and computer readable storage medium for similar film watching users |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103914518A (en) * | 2014-03-14 | 2014-07-09 | 小米科技有限责任公司 | Clustering method and clustering device |
| US20150169635A1 (en) * | 2009-09-03 | 2015-06-18 | Google Inc. | Grouping of image search results |
| CN105654039A (en) * | 2015-12-24 | 2016-06-08 | 小米科技有限责任公司 | Image processing method and device |
| US20160286272A1 (en) * | 2015-03-24 | 2016-09-29 | Fuji Xerox Co., Ltd. | User-profile generating apparatus, movie analyzing apparatus, movie reproducing apparatus, and non-transitory computer readable medium |
| CN106777007A (en) * | 2016-12-07 | 2017-05-31 | 北京奇虎科技有限公司 | Photograph album Classified optimization method, device and mobile terminal |
| CN108171252A (en) * | 2017-11-16 | 2018-06-15 | 柳州健鱼科技有限公司 | A kind of balanced image clustering method based on hierarchical cluster |
-
2018
- 2018-07-03 CN CN201810714996.0A patent/CN109063737A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20150169635A1 (en) * | 2009-09-03 | 2015-06-18 | Google Inc. | Grouping of image search results |
| CN103914518A (en) * | 2014-03-14 | 2014-07-09 | 小米科技有限责任公司 | Clustering method and clustering device |
| US20160286272A1 (en) * | 2015-03-24 | 2016-09-29 | Fuji Xerox Co., Ltd. | User-profile generating apparatus, movie analyzing apparatus, movie reproducing apparatus, and non-transitory computer readable medium |
| CN105654039A (en) * | 2015-12-24 | 2016-06-08 | 小米科技有限责任公司 | Image processing method and device |
| CN106777007A (en) * | 2016-12-07 | 2017-05-31 | 北京奇虎科技有限公司 | Photograph album Classified optimization method, device and mobile terminal |
| CN108171252A (en) * | 2017-11-16 | 2018-06-15 | 柳州健鱼科技有限公司 | A kind of balanced image clustering method based on hierarchical cluster |
Non-Patent Citations (3)
| Title |
|---|
| NAJVA IZADPANAH: "A DIVISIVE HIERARCHICAL CLUSTERING BASED METHOD FOR INDEXING IMAGE INFORMATION", 《SIGNAL & IMAGE PROCESSING : AN INTERNATIONAL JOURNAL (SIPIJ)》 * |
| 华秀丽等: "语义分析与词频统计相结合的中文文本相似度量方法研究", 《计算机应用研究》 * |
| 王全蕊等: "基于属性依赖关系和对象相关性的自然聚类算法", 《小型微型计算机系统》 * |
Cited By (22)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111382770A (en) * | 2018-12-29 | 2020-07-07 | 百度在线网络技术（北京）有限公司 | Picture clustering method, device, equipment and storage medium |
| CN110084267A (en) * | 2019-03-12 | 2019-08-02 | 北京旷视科技有限公司 | Portrait clustering method, device, electronic equipment and readable storage medium storing program for executing |
| US11232288B2 (en) | 2019-03-18 | 2022-01-25 | Shenzhen Sensetime Technology Co., Ltd. | Image clustering method and apparatus, electronic device, and storage medium |
| CN109800744A (en) * | 2019-03-18 | 2019-05-24 | 深圳市商汤科技有限公司 | Image clustering method and device, electronic equipment and storage medium |
| CN110245679A (en) * | 2019-05-08 | 2019-09-17 | 北京旷视科技有限公司 | Image clustering method, device, electronic equipment and computer readable storage medium |
| CN110245679B (en) * | 2019-05-08 | 2021-12-28 | 北京旷视科技有限公司 | Image clustering method and device, electronic equipment and computer readable storage medium |
| CN111062407A (en) * | 2019-10-15 | 2020-04-24 | 深圳市商汤科技有限公司 | Image processing method and device, electronic equipment and storage medium |
| CN110738258B (en) * | 2019-10-16 | 2022-04-29 | Oppo广东移动通信有限公司 | Image classification method and device and terminal equipment |
| CN110738258A (en) * | 2019-10-16 | 2020-01-31 | Oppo广东移动通信有限公司 | Image classification method and device and terminal equipment |
| CN111091152A (en) * | 2019-12-17 | 2020-05-01 | 上海云从汇临人工智能科技有限公司 | Image clustering method, system, device and machine readable medium |
| CN111368880A (en) * | 2020-02-17 | 2020-07-03 | 杭州海康威视系统技术有限公司 | Image clustering method and device, electronic equipment and computer readable storage medium |
| CN111368880B (en) * | 2020-02-17 | 2023-09-05 | 杭州海康威视系统技术有限公司 | Image clustering method, device, electronic equipment and computer readable storage medium |
| CN111310834A (en) * | 2020-02-19 | 2020-06-19 | 深圳市商汤科技有限公司 | Data processing method and device, processor, electronic equipment and storage medium |
| CN111401196A (en) * | 2020-03-10 | 2020-07-10 | 珠海全志科技股份有限公司 | Method, computer device and computer readable storage medium for self-adaptive face clustering in limited space |
| CN111460195A (en) * | 2020-03-26 | 2020-07-28 | Oppo广东移动通信有限公司 | Picture processing method and device, storage medium and electronic equipment |
| WO2021190165A1 (en) * | 2020-03-26 | 2021-09-30 | Oppo广东移动通信有限公司 | Picture processing method and apparatus, and storage medium and electronic device |
| CN111753114A (en) * | 2020-06-04 | 2020-10-09 | 北京字节跳动网络技术有限公司 | Image pre-labeling method and device and electronic equipment |
| CN111680753A (en) * | 2020-06-10 | 2020-09-18 | 创新奇智(上海)科技有限公司 | Data labeling method and device, electronic equipment and storage medium |
| CN112767410A (en) * | 2020-12-31 | 2021-05-07 | 北京航空航天大学 | MRI brain tissue clustering segmentation method |
| CN113052245A (en) * | 2021-03-30 | 2021-06-29 | 重庆紫光华山智安科技有限公司 | Image clustering method and device, electronic equipment and storage medium |
| CN113052245B (en) * | 2021-03-30 | 2023-08-25 | 重庆紫光华山智安科技有限公司 | Image clustering method and device, electronic equipment and storage medium |
| CN113378020A (en) * | 2021-06-08 | 2021-09-10 | 深圳Tcl新技术有限公司 | Acquisition method, device and computer readable storage medium for similar film watching users |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109063737A (en) | Image processing method, device, storage medium and mobile terminal | |
| CN108764372B (en) | Construction method and device, mobile terminal, the readable storage medium storing program for executing of data set | |
| CN108764208B (en) | Image processing method and device, storage medium and electronic equipment | |
| CN108830855B (en) | Full convolution network semantic segmentation method based on multi-scale low-level feature fusion | |
| CN108875821A (en) | The training method and device of disaggregated model, mobile terminal, readable storage medium storing program for executing | |
| CN108960290A (en) | Image processing method, device, computer readable storage medium and electronic equipment | |
| WO2019233393A1 (en) | Image processing method and apparatus, storage medium, and electronic device | |
| CN108764370A (en) | Image processing method, device, computer readable storage medium and computer equipment | |
| CN108777815A (en) | Method for processing video frequency and device, electronic equipment, computer readable storage medium | |
| CN110580428A (en) | image processing method, image processing device, computer-readable storage medium and electronic equipment | |
| CN108900769A (en) | Image processing method, device, mobile terminal and computer readable storage medium | |
| CN110276767A (en) | Image processing method and device, electronic equipment, computer readable storage medium | |
| CN108984657A (en) | Image recommendation method and apparatus, terminal, readable storage medium storing program for executing | |
| CN109002843A (en) | Image processing method and device, electronic equipment, computer readable storage medium | |
| CN107578390A (en) | A kind of method and device that image white balance correction is carried out using neutral net | |
| CN108810418A (en) | Image processing method, device, mobile terminal and computer readable storage medium | |
| CN108805103A (en) | Image processing method and device, electronic equipment, computer readable storage medium | |
| CN108764321B (en) | Image-recognizing method and device, electronic equipment, storage medium | |
| CN108875619A (en) | Method for processing video frequency and device, electronic equipment, computer readable storage medium | |
| CN108805198A (en) | Image processing method, device, computer readable storage medium and electronic equipment | |
| CN110473185A (en) | Image processing method and device, electronic equipment, computer readable storage medium | |
| CN108846351A (en) | Image processing method, device, electronic equipment and computer readable storage medium | |
| CN110233971B (en) | Shooting method, terminal and computer readable storage medium | |
| WO2019233260A1 (en) | Method and apparatus for pushing advertisement information, storage medium and electronic device | |
| CN109740721B (en) | Wheat ear counting method and device |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20181221 |
|
| RJ01 | Rejection of invention patent application after publication |